2005 REU Participants-Department of Biological Sciences - Carnegie Mellon University

2005 Participants

National Science Foundation
Mentored, Cutting Edge Research Experiences in the Molecular Biosciences
Research Experiences for Undergraduates (REU)

Ed BriercheckEd Briercheck, University of Toledo

Mentor: Dr. Brooke McCartney

The Use of Aptamers as a Novel Sensor of APC in Developmental Analysis of Drosophila melanogaster

Proteins have shown to be the key to unlocking the cellular mechanisms of organism development. Thus, the need for efficient and accurate methods for proper identification of proteins and their interactions in vivo has arisen. Using the fruit fly Drosophila melanogaster as a developmental model, our lab has determined that the Adenomatous polypolis coli (APC) protein family plays a significant role in development. To determine mechanisms of APC function it is crucial to be able to localize APC and other protein interactions within subcellular structures without disrupting the functionality of the endogenous protein. Satisfaction of these requirements may be found in oligonucleotide aptamers. Aptamers are short single-stranded DNA molecules which specifically bind to the target protein. In order to act as a sensor, the aptamers are fluorescently tagged by Watson and Crick base pairing interactions with fluorescently labeled peptide nucleic acids (PNA). The extreme size difference between the 100 base DNA and present sensor methods such as green fluorescent protein, suggest that aptamers may be less functionally disruptive to the target protein. To select for aptamers that specifically bind to Drosophila APC2, we expressed and purified the C-terminal half of the protein as a GST fusion, which was mixed with a synthetically produced randomized DNA pool generated by the systematic evolution of ligands by exponential enrichment (SELEX) protocol. We then subjected this mixture of APC2 protein with the DNA pool to nonequlibrium capillary electrophoresis of equilibrium mixtures (NECEEM), allowing us to definitively separate free DNA, APC2 bound DNA, and free APC2. Using this separation procedure, we isolated potential APC2 aptamers. We are currently amplifying these aptamers via PCR and will clone the optimum sequences. One technical hurdle of aptamer biosensors is the control of fluorescence. It is imperative that they only fluoresce when bound to the target. Otherwise significant background from unbound aptamers may mask those that are bound. The fluorescent difference between the two can be determined in vitro using a fluorimeter. Previously isolated aptamers for the protein thrombin verified this technique. A PNA-linked Dapoxyl fluorophore, showed significantly more fluorescence when bound to the protein than when free, making this a viable technique for future work on the aptamers for APC2.

Joanna CalderonJoanna Calderon, Pomona College

Mentor: Dr. Justin Crowley/Dr. Nathan Urban

Development of Imaging Quantum Dot-Based Neuronal Tracers

The goal of our project was to develop quantum dot-based fluorescent probes for use in anterograde and retrograde tracers in living neurons. Fluorescence-based imaging is an essential tool for the understanding of dynamic biological systems, however, the fluorescent probes currently in use have technological limitations that make them less than ideal for time-lapse, multicolor imaging experiments deep in the brain. These limitations include low quantum efficiency, photobleaching, and excitation/emission spectral ranges. The development of quantum dots (qdots), fluorescent semiconductor nanocrystals, has provided a fluourophore that combines very high quantum efficiency, resistance to photobleaching, and a broad excitation spectrum with a narrow emission spectrum (20nm). Thus, as a fluorophore, qdots are a vast improvement over traditional dyes. Therefore, this project aimed to develop and test coated qdots as anterograde and retrograde neuronal probes. Such a neuronal tracer could facilitate studies of time-lapse neuronal morphology in deep tissue. Qdots were injected intracellulary in vitro and extracellulary in vivo . Their utility as neuronal markers was assessed through a combination of in vivo , in vitro , and fixed tissue imaging. The initial experiments involved qdots coated with glutamic acid-AMP, AMP, PEG-COOH, PEG-5K, and streptavidin biotin. Preliminary results from in vitro intracellular qdot injections show that qdots do not alter the physiological properties of neurons, but labeling was limited to neuronal cell bodies ten minutes following the injection. Qdots injected in vivo , likewise, do not appear to harm neuronal viability within two days post-surgery. A time-course experiment showed that labeling of neurons around the injection site occurs within two hours following the injection. The qdots partially filled first, second, and third order dendrites near in vivo injection sites of glutamic acid-AMP, PEG-COOH, and PEG-5K coated qdots. Of these, PEG-COOH and PEG-5K coated qdots provided better dendritic labeling, however PEG-5K coated qdots labeled more completely. Cortical injections of AMP-coated qdots labeled the red nucleus of the thalamus, indicating a retrograde mechanism of transport. Streptavidin-biotin coated qdots almost exclusively labeled cell bodies. Future studies should examine the efficacy of slight modifications to PEG-5K qdots, both for in vivo and in vitro labeling.

Melissa FurlongMelissa Furlong, Duke University

Mentor: Dr. Eric Ahrens

Human Mitochondrial Ferritin as a Reporter for Magnetic Resonance Imaging

Human cytosolic H and L ferritin subunits can be used as reporters for transgene imaging using magnetic resonance imaging (MRI). When DNA from these subunits are transduced into a host tissue, the targeted cells produce ferritin shells that tend to accumulate iron from their surroundings. Consequently, a small iron-oxide nanocrystal is formed at the center of the ferritin, which produces MRI contrast of the cell in MRI. Human mitochondrial ferritin (Mtf) differs from cytosolic ferritin in that it is a homopolymer, using a single subunit rather than two subunits. In using the single subunit, the DNA sequence length that must be transduced is approximately one half that of the cytosolic construct, and this is advantageous technologically. The goals of the project included isolating, cloning, amplifying and delivering the Mtf gene into targeted host tissues using a replication defective adenovirus. Transduced tissue was examined using in vivo time-lapse MRI. As the Mtf DNA strands were GC-rich and notoriously difficult to separate, the Mtf was isolated and amplified using a PHusion kit designed specifically for GC-rich, triple bonded DNA sequences. TA cloning was first tried unsuccessfully to transform the sequence. Blunt ended ligations and transformations were then chosen as the optimal form of cloning. The remaining steps were not undertaken due to time restraints.

Sara HudsonSara Hudson, Willamette College

Mentor: Dr. John Woolford

Identifying Potential Binding Sites Between Ytm1, Nop7 and Erb1 in Saccharomyces cerevisiae 66S Pre-ribosomal Particles

Ribosomes are essential to cell metabolism, growth and reproduction. A mature eukaryotic ribosome is composed of a 60S (large) and 40S (small) subunit. Our research delves into the construction of the 60S subunit in Saccharomyces cerevisiae . The maturation process begins with the 90S subunit, composed of strands of RNA and proteins, which splits into the 66S subunit and the 43S subunit. The 66S subunit undergoes 6 different permutations before being released into the cytoplasm from the nucleolus, and interacts along the way with around 150 proteins not found in mature ribosomes. Woolford Lab currently researches two distinct transitory "neighborhoods" of several temporary proteins each: Nop7, Ytm1 and Erb1; and Rrs1, Rpf2p, Rpl11 and Rpl5. This summer, I focused on the proteins Nop7, Ytm1 and Erb1, which bind in some configuration on the 66S subunit. Previous work has shown that Nop7 and Ytm1 interact with Erb1 but not with each other, but did not show where and how the proteins bound to one another (Schleifman 2005). We wanted to clarify how and where interactions occur between these three proteins. Specifically, I looked for binding domains on Ytm1 and Erb1 with Aarti Sahasranaman, and binding domains of Nop7 and Erb1 on my own. Based on previous research on the human analogs Pes1 (Nop7) and Bop1 (Erb1), I hypothesized that the binding domain would be in the 21112 region, where several insertion mutations disturbed binding in vivo (Lapik et al 2004). I used two-hybrid analysis, Polymerase Chain Reaction (PCR), colony PCR, bacterial transformation and gel electrophoresis to find potential binding regions. Using PCR, I amplified certain regions of each gene, and tested the regions for their ability to bind in vivo through a two-hybrid screen. Neither experiment is finished at this time, although there are promising indications.

Lapik, Y., Fernandes, C., Lau, L., Pestov, D. 2004. Physical Interaction between Pes1 and Bop1 in Mammalian Ribosome Biogenesis. Molecular Cell, 15: 17-29.

Schleifman, E. 2005. A Microscaffold of Assembly Factors Necessary for Ribosome Biogenesis and Growth Control. BS Thesis, Carnegie Mellon University.

Claudia Lins_Claudia Lins, Montclair State

Mentor: Dr. Javier López

Mechanism of Activation of a Nonexonic Recursive Splice Site

Removal of introns from pre-mRNAs is required for expression of most genes. Many genes with important roles in development and disease have extremely long introns that exceed 20 kb. The mechanisms required for efficient expression of such genes are not well understood. One known strategy involves recursive splicing, which subdivides long introns with elements that function as 3' splice sites and regenerate a 5' splice site at the junction with the upstream exon. Most recursive splice sites are non-exonic. Two interesting questions concerning recursive splice sites are the following. How are these elements activated for splicing, and what directs their sequential activity as 3' and 5' splice sites? I have addressed these questions using the non-exonic recursive splice site RP3 from the Ultrabithorax gene of Drosophila . Phylogenetic comparisons have identified two conserved elements downstream of RP3 that suggest a hypothesis for the mechanism of activation. Element DE1 is located approximately 54 nt downstream of RP3 and contains a pseudo-5' splice site sequence. We hypothesize that interactions between U2AF bound at RP3 and U1 snRNP bound at DE1 lead to activation of RP3 as a 3' splice site, in the same way that 3' splice sites are recognized at standard exons. Element DE2 is located approximately 137 nt downstream of RP3 and contains a pseudo-3' splice site. We hypothesize that U1 snRNP bound at DE1 also interacts with U2AF bound at DE2, in the same way that it would across a short intron. In this way, DE1 and DE2 would define a pseudo-exon and pseudo-intron that are never spliced but that lead to activation of RP3 as a 3' splice site. I attempted to test this hypothesis by trying to mutate DE1 and DE2. A minigene ( Ubx .4F12) was used for these experiments in which the intron has been shortened from 52 kb to 2 kb. Previous work has shown that RP3 is functional in this minigene when assayed by transfection into the Drosophila SL2 cell line. I used PCR-mediated mutagenesis to delete or introduce point mutations at DE1 and DE2. However, I only succeeded at mutating DE2, not DE1. I believe the reason why the deletion of DE1 was not achieved was due to the remaining full-size plasmid of Ubx.4F12.RP (WT) following restriction enzyme digest tests. Meanwhile, another deletion DE"X" appeared, which is a larger deletion than DE2, approximately 500 bp larger. Not much is known about this deletion and how it will affect the recursive splicing machinery. At the moment, transfection into Drosophila SL2 cell line is being performed in order to analyze the function of RP3 in the minigene. Future plan is to use reverse transcription-PCR assays to examine the effects of these mutants on the activity of RP3 as a 3' splice site and as a regenerated 5' splice site.

Juliana LopezJuliana Lopez, Duke University

Mentor: Dr. Erik Thiessen

Determining the Domain Specificity of Rule Learning Within the Visual Domain of Infants

A fundamental part of language acquisition is learning the over-arching algebraic rules that govern phrase structure. Learning these rules makes recognizing and comprehending novel instances that follow such patterns possible. Is learning these types of rules specific to language, or can infants do it in other domains? The experiments done in the Infant Language and Learning Lab explore rule learning in the visual domain in order to explore whether rule learning is a "domain general" learning mechanism. We explored visual learning using simultaneously presented stimuli instead of sequentially presented stimuli to avoid memory overload for the 6 ½-7 ½ month-old infants. Two different sets of stimuli were generated using Macromedia Director and I-Movie. In the first experiment, infants were habituated to circles following a particular rule, ABA or ABC, during training trials. In the second experiment, infants were habituated to cats and dogs stimuli in either an ABA or ABB pattern. Once habituated, six novel rules sequences were used to test the infants for learning the particular rule; three sequences followed the rule they were accustomed to while three sequences violated the rule. We expected to see that an infant who learned the rule during the habituation trials would look longer at the novel items that violated the pattern they were exposed to. Instead, our experiment seems to show an unexpected preference for familiar sequences that follow the rule they saw during the habituation trials. However, our data still remains consistent with the idea that infants might still be learning visual rules in the same way they learn language rules. The fact that infants can differentiate patterns that follow and violate the sequence they were exposed to indicates that learning is taking place. If this is true, it suggests that language learning isn't "special," but instead, it follows the same mechanism that is used in everyday learning.

Chris SternChris Stern, St. Vincent College

Mentor: Dr. Alison Barth

Timing of Gene Expression Using fos-Timer Reporter Construct

C -fos is an activity dependent early-gene that functions as a transcription factor in central nervous system neurons. Transgenic mice carrying the green fluorescent protein (GFP) fused in frame with c-fos ( fosGFP ) have allowed in vivo identification and characterization of specific neurons involved in particular functions. While fosGFP has proven to be useful in this context, it provides no information regarding the time frame of c- fos expression. A variant of a red fluorescent protein, Timer, emits fluorescence that changes from green, to yellow, and finally to red over time. Hence, a fosTimer construct has been created to determine when and for how long c- fos has been expressed in individual neurons. However, Timer must exist as a homotetramer in order to fluoresce. Since functional Fos exists as a heterodimer with a Jun family member, fosTimer generates a multi-subunit protein complex that results in decreased cell viability. We hypothesized that mutating the leucine-zipper domain involved in Fos-Jun dimerization would reduce potentially toxic effects to the cell. The goals of this project were to generate and test in vitro this modified fosTimer construct. First, site-directed mutagenesis was performed on fosTimer and the products were sequenced in order to screen for desired constructs. To determine whether mutated Timer behaved as expected, cell line A549 was transfected with fosTimer , mutated fosTimer ( MfosTimer ), and fosGFP . Forskolin was added to induce Fos expression and cells were subsequently fixed 0, 6, 12, 24, 36, and 48 hours after induction. Cell toxicity was also analyzed at 48 hours using an MTT toxicity assay. The data indicate that while MfosTimer resulted in decreased toxicity as compared to fosTimer , cells transfected with MfosTimer displayed only green fluorescence that did not mature to red. This result may indicate that MfosTimer products (containing non-functional, mutated Fos) are degraded before they can mature. The data also indicate that Timer from cells transfected with fosTimer behaved as expected. Interestingly, cells transfected with fosTimer also emitted fluorescence longer than cells transfected with fosGFP , suggesting that Timer will indeed constitute a powerful tool for the temporal analysis of immediate early gene expression in neurons.

Olga SvetnikovOlga Svetnikov, Caldwell College

Mentor: Dr. Frederick Lanni

Development of a Photo-Activated Inhibitor of GTPase Signaling to the Actin Cytoskeleton

Cdc42 is a Rho-family GTPase that activates downstream effectors, producing specific actin cytoskeleton structures and behaviors. Cdc42 in its activated form (Cdc42(GTP)) binds a number of proteins, one of which is WASP (Wiskott-Aldrich Syndrome Protein), that ultimately result in actin polymerization, filopod formation and cell polarization. Because cytoskeletal processes occur on a rapid timescale, many experimental interventions such as genetic modification are often too slow to be useful. To detect signal transduction in GTPase pathways affecting the actin cytoskeleton, I have utilized CBD, the Cdc42-binding domain of WASP, which, when introduced into cells in sufficient amounts, inhibits the Cdc42 signaling pathway by out-competing native effectors. By chemically caging CBD with NVOC-Cl, a light-sensitive organic compound that reacts with lysine residues, I tested for inactivation of the domain's inhibitory activity, and analyzed its reactivation via UV light. Pull-down assays gave inconclusive results in vitro , while live cell experiments were more promising. To assay GTPase signaling in vivo , Swiss 3T3 fibroblasts on the edge of a wound were microinjected with native, caged and uncaged CBD. I have found that, in vivo , caged CBD shows a lower inhibitory effect than non-caged CBD, as expected. Work continues to optimize dosage, caging efficiency and uncaging conditions.

Wesley VosburgWesley Vosburg, Washington and Jefferson College

Mentor: Dr. Gordon Rule

An Investigation into the Dynamics of Class Alpha Glutathione Transferases

Glutathione transferases (GSTs) play an important role in cellular defense against various chemicals and may also serve as binding proteins which participate in intracellular organic ligand transport. The defensive role of GSTs is one of the causes for tumor cell resistance to cytostatic drugs. Studying different isozymes and mutants of these enzymes will lead to an understanding of the significance that the enzymes actually have. Previous studies have shown that the active site of the enzyme is covered by the C terminal alpha helix. Crystallographic information had suggested that the alpha helix structure was not defined without the presence of substrate, but with the use of NMR it has been shown that residues within the alpha helix are present in the helical structure without the presence of substrate. The goals of this project are to introduce mutations into a specific amino acid in the GST A1-1 enzyme that affects the stability of the C terminal helix and interpret the effect which they have on the enzyme's binding affinity. Methods which will be used to accomplish this goal include bacterial transformation, protein purification, enzyme assays and isothermal titration calorimetry. Upon completion of the project there is a further understanding of the role which specific residues exhibit in the function of A1 glutathione transferase.

Lauren WilliamsLauren Williams, Immaculata College

Mentor: Dr. Tina Lee

The Effects of N-ethylmaleimide on Endoplasmic Reticulum Morphology and Vesicle Budding Activity in vivo

The endoplasmic reticulum (ER) is a single membrane system with continuous intralumenal space, and it has three distinct domains: smooth, rough, and transitional. This experiment focuses on transitional ER, which is involved in packaging of proteins for transport to the Golgi apparatus and is enriched in proteins for the process. A protein of interest is the p97, an ATPase with a cysteine residue that is critical to enzyme activity. The complex is involved in ER network formation, ER exit site formation, and coat protein complex (COP-II) mediated vesicle budding. The goal of this experiment was to determine the effects of N-ethylmaleimide (NEM) on COP-II mediated vesicle budding and ER network density. To investigate this, the catalytic site on p97 was obstructed by the inhibitor NEM, which alkylates the sulfhydryl group on cysteine residues. When the ATPase was made catalytically inactive, a perturbation of the ER network and a decrease in COP-II resulted. HeLa cells were transiently transfected with Sec61b fusion protein to express for Green Fluorescent Protein (GFP). Cells were then treated with NEM over time and immediately placed into fix. The cells were stained with COP-II through immunofluorensce. Transfected cells were then examined using fluorescent and confocal microscopy. Results showed that the inhibitor NEM dramatically reduced the amount of COP-II activity over time in intact cells, indicating a critical intermediary in the COP-II vesicle budding mechanism. Data also displayed a decrease in the density of ER network following treatment with NEM. Confocal microscopy clearly showed the perturbation of the ER network increasing over time with NEM treatment. Both COP-II mediated vesicle formation and ER network formation are NEM-sensitive processes. An intermediary complex with a cysteine residue has been implicated in budding activity and exit site formation.

Please send inquiries about our Research Experiences for Undergraduates program to bio-reu@andrew.cmu.edu.